Vol. 9, 505-512, July 1998 Cell Growth & Differentiation 5C5

Epidermal Growth Factor-induced Growth Inhibition RequiresStati Activation1

Jacqueline F. Bromberg, Zhen Fan, Carol Brown,John Mendelsohn, and James E. Damell, Jr.2Laboratory of Cell Biology, The Rockefeller University, New York, NewYork 10021-6399 [J. F. B., J. E D.]; Department of ClinicalInvestigation, The University of Texas, M. D. Anderson Cancer Center,Houston, Texas 77030 [Z. F., J. M.]; and Departments of Surgery andGynecology, Memorial Sloan Kettering Cancer Center, New York, NewYork 10021 [C. B.J

AbstractEpidermal growth factor (EGF) is a mitogen for mostepithellal cells. Paradoxically, the growth of some

cultured cell lines, containing high numbers of EGFreceptors, are inhibited by EGF. Here we demonstratethat growth inhibition by EGF In several cell lines

correlates with the activation of the signal transducer

and activator of transcription (Stat) 1. In contrast, in

normal fibroblasts and several cell lines that aregrowth stimulated by EGF, we observed no or verytransient activation of Stati . A causal associationbetween Stati activation by EGF and growth inhibition

was suggested by the expression of a dominant-negative Stati in A431 cells, resulting in the loss of

Stati DNA binding and concomitant resistance togrowth inhibition by EGF. We conclude that, In the cellsexamined, EGF-lnduced arrest of growth requiresactivated Stati.

Introduction

EGF3 stimulates growth of many normal and malignant celllines (1). Nevertheless, a number of tumor cell lines that

contain high numbers of EGF receptors are growth inhibitedby nanomolar amounts of EGF (2-7). This paradoxical phe-nomenon has been best characterized in A431 cells, a linederived from a vulvar carcinoma (2, 4). These cells contain ahigh number of EGFRs, specifically erbl (4). EGF induces ablock in the G1 phase of the cell cycle with a correspondinginhibition of CDK2 activity because of the increased produc-

Received 4/7/98; revised 5/21/98; accepted 5/21/98.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to mdi-

sate this fact.1 This work was supported by NIH grants A132489 and A134420 (toJ. E. D.). J. F. B. was supported by NIH Training Grant CA09207, aHoward Hughes Postdoctoral Fellowship for Physicians, and NIH K08Grant CA67950.2 To whom requests for reprints should be addressed, at Laboratory ofCell Biology, The Rockefeller University, 1230 YorkAvenue, New York, NY10021-6399. Phone: (212) 327-8791 ; Fax: (212)327-8801.3 The abbreviations used are: EGF, epidermal growth factor� EGFR, EGFreceptor; CDK. cyclin-dependent kinase; Stat, signal transducers andactivators oftranscnption; MARK, mitogen-activated proteln kinase; TCA,trichloroacetic acid; EMSA, electrophoretic mobility shift analysis; MBP,myelin basic protein; BCS, bovine calf serum.

tion of the CDK2 inhibitor CIP1/WAFI/p21 (8, 9). The mech-anisms leading to increases in p21 production in A431 cellsor in other cell types have not been determined, although anumber of molecules have been implicated, including p53(10, 11), IRF1 (12), and Stati (9).

Exposure to EGF activates the EGF signal transductionpathway through homo- and heterodimenzation of differentreceptors in the EGF receptor family, followed by transphos-phorylation of the receptor molecules on several different

tyrosine residues. Numerous proteins containing SH2 do-mains then bind to the tyrosine phosphates on the receptor(1 , 13, 1 4). Once bound to the receptor, many of theseproteins also become tyrosine phosphorylated, presumably

by the kinase domain ofthe EGFR, and mediate a cascade ofbiochemical events. Some of these proteins that mediateEGFR signaling include syp, vav, shc, grb2, PLC’y, P13Kp85,Statl , and Stat3 (14-20). Why growth stimulation occurs inmost cases of EGF treatment and growth inhibition in other

cases is not known.To better understand EGF-mediated growth inhibition, we

examined Stat protein activation in cell lines that were eithergrowth stimulated or inhibited by EGF. Stats are latent tran-scription factors that upon ligand binding become tyrosinephosphorylated, form dimers, translocate to the nucleus,

recognize specific Stat DNA-binding elements, and inducetranscription (21 , 22). Statl and Stat3 are tyrosine phospho-rylated after treatment of A431 cells with EGF (1 6, 17). Thephosphorylation event may be mediated by the receptortyrosine kinase (EGFR) or by Jakl (23, 24), both of which are

activated in EGF-treated cells. We evaluated Stat phospho-rylation in cell lines containing either high numbers (2 x 106)of EGFR, such as the A431 cell line (4) and A431 variant cellsthat are not growth inhibited by EGF (8), moderate numbers(1 x 10�), such as the cervical carcinoma cell lines SiHa and

Me180 (25), or low numbers (4 x 10k), as seen in primaryfibroblasts (26, 27). In all cases examined, Stati activation by

EGF correlated with growth inhibition by this ligand. Whengrowth stimulation was observed, no Stati phosphorylationwas detected. However, another EGF-activated pathway,the MAPK pathway, was enhanced by EGF in all of thedescribed cell lines.

To demonstrate a causal association between Statl acti-vation and EGF induced growth inhibition, we introduced intoA431 cells a Stati molecule lacking its terminal tyrosine 701.This molecule presumably binds to the phosphorylatedEGFR through its SH2 domain but cannot be phosphorylatedon residue 701 and subsequently interferes with wild-typeStati phosphorylation. In A431 subclones containing thisconstruct, EGF stimulation did not lead to phosphorylation ofwild-type Statl , nor did growth inhibition occur. Thus, inA431 cells, Stati activation appears to be required for EGF-induced growth arrest.

A. B.A431-P

days

Fig. 1 . Growth of A431 -P cells and Stati activation. In A, A431 -P cellswere grown in DMEM with 0.5% BCS and treated with EGF (E; 20 nM) or

IFN-y (y; 1 5 ng/ml) for 1 h. Nuclear extracts were made from these cells,and -10 �g of protein were incubated with a 32P-labeled, high-affinityStati and Stat3 DNA binding site (m67, dGATTTCCCGTAAATCAT) andanalyzed by EMSA. The addition of Stati antisera (Si) resulted in super-shifting the Statl/DNA complex to a high mobility form, thus specifyingthe complex as Stati containing. Preimmune serum or Stat3 antisera doesnot supershift these complexes (data not shown). In B, A431 -P cells wereplated at low density (20-30%) in DMEM with 0.5% BCS. After 6 h, thecells were adherent, and EGF (E, #{149}; 20 nM) or IFN-y (� #{149};1 5 ng/ml) wasadded. The average cell number from triplicate samples was determineddaily (by Coulter counter). #{149},untreated.

5� Stati and EGF-induced Growth Arrest

4 J. Bromberg and J. E. Damell, Jr., unpublished observations.5 z. Fan and J. Mendelsohn, unpublished observations.

A431-P

Antibody Si Si

Ligand - E E I I

Stat!

ResultsEffects of IFN-y and EGF on Growth and Stat Activation

in A431 Cells. A431 cells are growth inhibited by EGF andIFN--y (4, 9), and both EGF and IFN-y activate the phospho-

rylation of Stati protein (1 7-1 9, 28). The growth and activa-tion of Stati in A431 cells in response to either IFN-y or EGFwas examined. Activated Stati is known to be tyrosine phos-phorylated, homodimerize, binds DNA, and mediates tran-scriptional activation (29, 30). A431 cells were growth inhib-

ted by both ligands, and Stati was activated as tested byDNA binding in response to treatment with IFN--y and EGF(Fig. 1). There was a very weak activation by EGF of Stat3 in

these cells (as seen in Figs. 1A and 4C; the slower migrating

band above the Statl homodimer is a Stati :3 heterodimer;

data not shown). Stat3 activation by EGF was described

earlier after 15 mm of EGF treatment (1 7, 31). The activatedStat3 disappears within 60 mm, the time of the experiment in

Fig. 1.Effects of IFN-y and EGF on Growth and Stati Activa-

tion in Two Cervical Carcinoma Cell Lines: Me180 and

SiHa. We next measured the growth of two cervical carci-noma cell lines in response to EGF and IFN-y. These cell

lines contain about 1 0-fold fewer EGFRs (1 x 10�) than found

in A431 cells (25). Me180 cells are growth inhibited by EGF,whereas SiHa cells are slightly growth stimulated by EGF

(Fig. 2A). Both cell types were growth inhibited by treatmentwith IFN-y (Fig. 2A). EGF stimulation of Mel 80 and SiHa cellsresulted in the phosphorylation of the EGFR and induction ofMAPK activity (Fig. 2, B and C). Thus, the EGFR can be

phosphorylated in both cell types, and conservation of anEGF-activated pathway is maintained.

These cell types differ, however, in their ability to stimulate

Statl phosphorylation in response to EGF. In the Mel 80 cellline, Statl phosphorylation (confirmed by DNA binding and

Statl antibody interaction) was observed after treatment withEGF and was maintained for -45 mm (Fig. 2D, Lanes 4 and6; data not shown). In contrast no Statl phosphorylation wasobserved in the SiHa cell line after treatment with EGF (Fig.

2D, Lane 5). Stat3 activation was not observed in either cellline. The Statl protein in the SiHa cell line could be activated

because treatment of these cells with IFN-’y resulted in Statlphosphorylation (Fig. 2D, Lanes 1 1 and 13). Furthermore, thephosphorylated Statl protein that binds DNA was transcrip-tionally competent because it induced IRF1 gene transcrip-tion (Fig. 2E). Transcriptional induction of IRF1 requires Stati(29, 32, 33)#{149}4

Effects of IFN-y and EGF on Growth and Stati Activa-

tion of Primary Fibroblasts. Primary fibroblasts typically

have 4 x 10�-l x l0� EGFRs and are growth stimulated by

EGF (26, 27, 34, 35). Therefore, we determined the growthresponse of BUD8 cells (diploid, nontransformed human fi-broblasts derived from skin) to EGF and IFN-y. These cells

were growth inhibited by IFN-y and growth stimulated by

EGF, as determined by cell number and tritiated thymidineuptake (Fig. 3, A and B). Nuclear extracts from EGF- andIFN-y-treated BUD8 cells showed Statl DNA binding corn-

plexes after IFN-y treatment but no Statl activation by EGF(Fig. 3C). Thus in this cell type, as shown in SiHa cells whereEGF leads to growth stimulation, no Statl activation wasseen. MAPK activity, however, was enhanced by EGF inthese cells (data not shown).

An A431 -derived Cell Line Is Growth Stimulated by

EGF, which Correlates with Only Transient Stati Activa-

tion. To further examine the mechanism of EGF-induced cell

cycle arrest in cell types with identical genetic backgrounds,we used subclones ofA43l -P (parental) cells (8). The A431 -C

subclone contains approximately one-half the number ofEGFRs compared with A43l-P cells, as determined by 125I

labeled EGFR saturation binding assay and Western blotanalysis (Ref. 8; Fig. 4D, bottom panel, Lane 1).� Growth inresponse to EGF was compared by tritiated thymidine up-take analysis (Fig. 4A). Serum alone resulted in a modest

increase in uptake over 72 h in both A431-P and A43l-C

lines; however, EGF treatment resulted in a marked decreasein uptake over time in the A431 -P line and only a transientimmediate decrease in uptake in the A431-C line, with a

subsequent increase between 12 and 72 h.EGF is known to induce a G1 cell cycle block in A43l -P

cells with an inhibition of CDK2 by the inhibitor p2l/CIP1/

WAF1 (8). We examined CDK2 activity in A43l -P and in

A431 -C cells in response to EGF; the expected inhibition wasobserved in A431 -P cells, but only transient inhibition oc-

curred in A43l -C cells that lasted less than 1 2 h, consistent

with the thymidine uptake results (Fig. 4B and data notshown). Stat binding activity was determined in both celllines in response to EGF at several time points (Fig. 4C).

Statl was activated for at least 48 h in the A431 -P line but

only briefly in the A43l -C line. Why Statl phosphorylation

was maintained in the parental line but only transiently acti-

days days

B. EGFR IP

MelSO Silla

-Ey -ET

EGFRPY �

C. MAPK IP MBP Kinase

MelSO Sills- yE -

MBP �

Ligand- E � - E

CeIILIneMMM S S S

LRF1�1S aGAPDH

Cell Growth & Differentiation 507

A. Growth Curves

MelSO Silla

D. Gel Shift E. Northern

SIFA(3:3)

SLFB(3:1) -

SIFC(1:1)

CeilLine A MS MS MS MS MS MS

LigandE- - EEEEEE’Y’Y’y’y

Antibody S1S1S3S3 S1S1

Fig. 2. Response of Me180 and SiHa cervical carcinoma cell lines to EGF and IFN-y. In A, Me180 and SiHa cells were plated at low density in RPMI with0.5% BCS. After 6 h, IFN-y � C; 1 5 nglml) or EGF (E, A; 20 nM) was added to the media. The average cell number from triplicate samples was determineddaily (by Coulter counter). In B, whole-cell extracts were isolated from Mel 80 and SiHa cells treated with IFN-y and EGF for 1 h. One hundred �g of totalprotein from each extract were immunoprecipitated with antisera to the EGFR, and the immunoprecipitated products were analyzed on an 8%polyacrylamide gel, transferred to nitrocellulose, and probed with an antiphosphotyrosine antibody, PY2O. - , untreated; ‘y, IFN-y treated; E, EGF treated.In C, whole-cell extracts were isolated from Mel 80 and SiHa cells treated with IFN-y and EGF for 1 h. Two hundred �g of total protein from each extractwere immunoprecipitated with antisera to ERK1 and ERK2, and the immunoprecipitated product was incubated with [y-32P]ATP and MBP. The radiolabeledMBP was analyzed by electrophoresis. Equal amounts of ERK1 and ERK2 were immunoprecipitated and loaded (data not shown). - , untreated; � IFN-ytreated; E, EGF treated. In D, nuclear extracts from A431 (A), Mel 80 (M), and SiHa (5) cell types treated with EGF (E) or IFN-y (y) for 1 5 mm were incubatedwith radiolabeled m67 probe and analyzed on a nondenaturing 4% polyacrylamide gel. Supershifting with Statl antisera (Si) but not Stat3 antisera (53)specifies these complexes as Statl containing (SS, supershift). SIFA is a Stat3 homodimer, SIFB a Statl/Stat3 heterodimer, and SIBC a Statl homodimer.In E, total RNAfrom Me180 (M) and SiHa(S) cells treated with EGF (E) or IFN-y(y)for4 h was resolved on a denatunng formaldehyde agarose gel, transferredto a nylon membrane, probed with radiolabeled IRF1 probe, and reprobed with glyceraldehyde-3-phosphate dehydrogenase (GAPD!-1) as a loading control.- , untreated.

vated in the A431 -C line is unknown. The two cell lines have,within a factor of two, similar numbers of EGFRs, yet quali-tative differences in the receptors are not known to exist.

Receptor internalization and destruction is thought to be

important in decreasing EGF-induced signal transduction

pathways (34, 36). Therefore, we examined the fate of theEGF receptor at several time points after exposure of both

cell lines to EGF (Fig. 40). Samples of cell extracts were

precipitated with anti-EGF receptor antiserum, and precipi-

tates were tested for phosphotyrosine on the EGF receptor(resolved by electrophoresis on acrylamide gels). In theA43l -C line, the tyrosine phosphorylated EGFR (Fig. 4D,

center panel) is maintained for 12 h but declines thereafter

compared with the A431 -P line, where the phosphorylated

A. Growth Curves

C.

15

10

5

EG}

.�-

“0 � � 3 � . 6days

B.

I7500

5000

2500

o� U

I EGF IFNy

Fig. 3. Response of primary fibroblasts to IFN-y and EGF. In A, BUD8cells were plated in DMEM with 10% FCS (1 x lO5ceIIs/six-weII dish) andtreated with 15 ng/mI of IFN-�y (‘y) or 20 n�.i EGF. The average cell numberfrom triplicate samples was determined after 3 and 6 days (by Coultercounter). In B, BUD8 cells were plated at 5 x l0� cells/six-well dish andtreated as described above. After 12 h, cells were pulsed with [�H)thymi-dine (1 pCVmI; 70 CVmmoI) for 1 h, after which time cells were washedand collected; TCA-precipitable counts were determined for each sample.U, untreated; El, EGF; 0, IFN-y. In C, nuclear extracts from untreated (-),EGF (E)-, or IFNy (‘y)-treated BUD8 cells were analyzed by EMSA. Cellswere treated for 30 mm with their respective ligands.

5� Stati and EGF-induced Growth Arrest

3-H Thymidine Uptake

receptor is maintained at high levels for 48 h. The total EGFRcontent of the two cell lines (determined by Western blots,Fig. 4D, bottom panel) showed a dramatically greater and

more rapid loss in A431 -C cells than in the parental cells. Thisprolonged activated EGFR may therefore be related to pro-longed Stati activation in A431 -P cells. Finally, we also ex-amined MAPK activity in both cell types and found EGF-induced MAPK activity in both with perhaps slightmaintenance of MAPK activity in the A431 -P line after 48 h(Fig. 4E), although growth inhibition was in effect at this time.

Blocking Stati Phosphorylation in A431 Cells RelievesEGF-induced Growth Arrest. To attempt to determinewhether activation of Stati was required for growth arrest ofthe parental A431 cells, we introduced a mutant epitope-tagged Stati protein (Y701 F) by transfection and selectedstable cell lines expressing the mutant protein. The exchangeof Y701 for F produces a molecule that should still bind to thephosphorylated EGFR through its SH2 domain but cannot bephosphorylated. Such a molecule might block wild-typeStati from binding to the EGFR (Fig. 5A). A similar result wasobtained with a Stat2 mutant, which decreased wild-typeStat2 phosphorylation at the IFN-a receptor (37), with a V-F

mutant in Stat3 in interleukin 6 signaling (38, 39) and with aV-F mutant in Stati in IFN’y signaling (38). Characterization ofa representative subclone, A5, that expressed the epitope-tagged mutant V701 F protein is depicted in Fig. 5. We ex-amined EGFR expression and phosphorylation in responseto EGF in these clones (Fig. 5C). Approximately equalamounts of total EGFR was found in untreated cells, butthere was a significant decrease in both the total amount ofreceptor and phosphorylated receptor after EGF stimulationin A5 cells within 20 mm of EGF treatment. The basis of rapidturnover of EGFR upon ligand addition in the cells expressingStati -V-701 F is unknown.

We examined Stati phosphorylation and binding to DNAby gel shift analysis in AP and A5 cells in response to EGFand IFN7 (Fig. 5D). Stati activation was not observed in AScells after EGF treatment but was observed after treatmentwith IFN-1. Thus, in some manner, StatlV-F can blockEGFR-mediated Stati phosphorylation but does not signifi-cantly lower IFN-y activation of Stati . The EGFR in both APand A5 cell types was capable of activating the MAPK path-way (Fig. SE), despite the relative quick decrease of phos-phorylated EGFR in the A5 cell type. Finally, we analyzed thegrowth response of the Stati V-F-containing clones to EGF

and IFN-y. AS cells are not growth inhibited by EGF, but theirgrowth was slowed by IFN-’y. In contrast, the parentalA431 -P cells are growth inhibited by both (Fig. 519.

DiscussionActivation of Stati by IFN-a or IFN-’y leads to growth arrestin many cell types, and this appears to require wild-typeStati . Stati -deficient cells or cells containing a transcnp-tionally defective Stati are no longer growth inhibited bylFNs (9, 40). Prolonging the half-life of phosphorylated Statialso promotes growth inhibition (41). Constitutively phospho-rylated Stati , as seen in achondroplastic cells because of aconstitutively activated fibroblast growth factor receptor, hasalso been reported to correlate with growth arrest, possiblydue to the increased presence of the cell cycle inhibitor p21(42). Thus, activation of Statl , and especially maintenance ofits activated state, can lead to growth inhibition.

EGF can induce phosphorylation of Statl and Stat3 (17).However, in the cell types examined in this report, we see amore pronounced and prolonged activation of Statl by EGF.Specifically, in response to EGF, Stat3 is phosphorylated inA431 -P cells only weakly and briefly, whereas Statl is acti-vated for at least 48 h. In Mel 80 cells, only Statl is phos-phorylated in response to EGF. Thus, in Mel 80 and A431 -P

cell lines, there is a clear correlation between EGF-inducedgrowth arrest and Statl activation. Furthermore, in a corre-spondingly matched cell line A431 -C and in SiHa cells as wellas BUD8 cells, where EGF does not induce growth arrest,either no or transient Statl activation was seen.

To demonstrate a requirement for EGF-dependent growtharrest and Statl activation, we introduced into A43l cells aStatl molecule that cannot be phosphorylated on tyrosine701 . Putative Stat docking sites on the intracellular domain of

the EGFR have been determined (43). A Statl moleculewhich cannot be phosphorylated on V-701 might dock onto

A431-P

10ooo�

E

5O00�

A431-C

E10000

5000

kZZ21��0 4 12 i4 ‘is 72

EGF (his)

A431-C

Hi

0 412244872EGF (lirs)

B. CDK2 IP Hi Kinase Assay

A431-P

�

0 4 12244872

EGF (lirs)

C. Gel Shift Assay: Stati binding

A43i-P

SIFC

EGF (hrs)

D. EGFR Immunoprecipitation

A431-C

� �..

0 1 1224 48

EGF (hrs)

A431-P A43i-C

EGFQIrS) . 1 4122448 - 1 4122448

V�n ‘�

EGFR � .�..,

Western

Cell Growth & Differentiation 509

A. 3-H Thymidine Uptake

Fig. 4. A431 -P versus A431 -C cell lines. In A,A431-P and A431-C cells were grown in DMEMwith 0.5% BCS in six-well dishes at 1 x 1O5cells/well. EGF (C ; 20 nM) was added to each well, andat the various time points indicated, rHlthymidine(1 �Ci/ml; 70 CVmmol) was added. After 30 mm,the cells were washed and collected, and TCA-precipitable counts were determined for eachsample. Each time point represents the averageof three independent experiments. U, untreated.In B, extracts(200 i.�g ofprotein)from A431-P andA431 -C cells treated with EGF (20 nM) for theindicated times were immunoprecipitated withanti-CDK2 antisera, followed by incubation of theimmunoprecipitated beads with a reaction solu-tion containing the reagents required for assayinghistone Hl phosphorylation (8). Radiolabeled hi-stone Hi protein was resolved on a 12% SDS-PAGE. Equal amounts of CDK2 were immunopre-cipitated and loaded (data not shown). In C,nuclear extracts from A431 -P and A43i -C cellstreated with EGF (20 nM) for the indicated timeswere incubated with a radiolabeled m67 probeand analyzed on a nondenatunng 4% polyacryl-amide gel, as described in “Materials and Meth-ods.” SIFC corresponds to a Statl homodimer. InD, extracts (150 ;.Lg of protein) from A431-P andA43i -C cells treated with EGF (20 nM) for theindicated times were immunoprecipitated withanti-EGFR antisera. - , untreated. Immunopre-cipitated protein was resolved on a 7% SDS-polyacrylamide gel, analyzed by Westem blotanalysis with antisera against phosphotyrosmne,stripped, and reprobed with antisera against theEGFR. In E, extracts (200 �g of protein) fromA431 -P and A431 -C cells treated with EGF (20nM) for the indicated times were immunoprecipi-tated with antisera against ERK1 and ERK2, fol-lowed by assaying MBP phosphorylation. Equalamounts of ERK1 and ERK2 were immunopre-cipitated and loaded (data not shown). -, un-treated.

o 12244872

EGF (hrs)

E. MAPK IP MBP Phosphorylation

A431-P A431-C

EGF(hTS) - 148 - 1 48#{149}.�.I. �*

the receptor but not be released as efficiently as when tyro-sine phosphorylation occurred, thus decreasing the accessof wild-type protein to the receptor. A similar Stat3 domi-

nant-negative (Stat3V7O5F) has been shown to block inter-

leukin 6-mediated Stat3 activation through the gpl3o recep-tor chain (38, 39, 44), as has a Stat2 mutation for the IFN-areceptor (37). However, activation of Statl by IFN-y was not

impeded by the introduction of the Statl Y70l F, whereas the

EGF activation of the endogenous wild-type Stat was. It istherefore possible that the IFN-’y receptor-Statl interactionmay be more transient, allowing wild-type Statl adequate

access to the receptor. In contrast, the release of Statl fromthe EGFR might require phosphorylation of the substrate.Possibly such blocked receptors are also targeted for deg-

A. DominantfNegative Stat!

Statkx Y-F

701 Flag

B.M2IP SlIP

APASAPA5

� Si Western

D. Gel Shift Assay

A-P

Antibody Si SiLigand - E E I I

A-S

Si

-1 ? E

C. EGFR IPA-P A-S

Li� - E - E

I P PY2O Western

It I I EGFR Western

E. MAPK IP MBP plies

- E

AP

A5

F. Growth Curves

. 55

SWC�$41

A43!-P A43!-5

days246

days

510 Statl and EGF-induced Growth Arrest

Fig. 5. StatiY-F blocks endogenous Stati activation by EGF. A, StatlY-F/FIag mutant used to block Stati activation by EGF. In B, whole-cell extracts(1 00 �g of protein) from A431 -P (.4P) and an exemplary clone expressing the dominant-negative protein A431 -5 (/15) were immunoprecipitated with anti-flagantisera (M2IP) or anti-Stati antisera (SlIP). Immunoprecipitated proteins were resolved on an 8% SDS-polyacrylamide gel, transferred to nitrocellulose,and probed with antisera against Stati . Proteins were visualized using ECL In C, whole-cell extracts (1 00 �g of protein) from A431 -P (AP) and A43i -5 (.45)were immunoprecipitated with anti-EGFR antisera, resolved on a 7% SDS-polyacrylamide gel, and probed with anti-phosphotyrosine antisera (PY2O). Theblot was stripped and reprobed with anti-EGFR antibody. Proteins were visualized using ECL - , untreated. In D, nuclear extracts (1 0 �ig of protein) fromA431 -P (,SIP) and A431-5 (.45) cells treated with EGF (5) or IFN-’y (‘y) were incubated with radiolabeled m67 probe and analyzed on a nondenaturing 4%polyacrylamide gel, as described in “Materials and Methods.” SIFC corresponds to a Statl homodimer as determined by supershifting (55) this complexwith antisera to Statl (Si). In E, extracts (200 �g of protein) from A43i -P (AP) and A431 -5 (.45) cells treated with EGF (20 nM) for 30 mm wereimmunoprecipitated with antisera against ERK1 and ERK2, followed by assaying MBP phosphorylation. Equal amounts of ERK1 and ERK2 wereimmunoprecipitated and loaded (data not shown). In F, A431 -P and A431 -5 cells were plated at low density in DMEM with 0.5% BCS. After 6 h, the cellswere adherent, and EGF (E, 20 nM) or IFN-y (y, 15 ng/mI) was added. The average cell number from triplicate samples was determined daily (by Coultercounter). -, untreated.

radation, explaining the rapid loss of EGFRs in EGF-treatedcells bearing the Stat V70l F mutant.

Whatever the molecular basis for the effect of the intro-duced Stat Y701 F, there was interruption of EGF-induced

phosphorylation of Statl with concomitant release from

EGF-induced growth restraint. The EGFRs, however, re-

mained functional insofar as MAP kinase activation was con-

cerned. The results of these studies, coupled with earlier

work on the requirement of Statl in growth restraint inducedby IFNs (9, 40), strongly suggest a general role of Statl inbalancing positive growth signals.

Materials and MethodsCell Lines, Culture Conditions, and Ligands. A431 , Me180, SiHa, andBUD8 cell lines were obtained from the American Type Culture Collection.The A431 -C cell line was isolated as described (8). A431 -P and A431 -C

Cell Growth & Differentiation 511

cells were grown in DMEM with 0.5% or 10% Cosmic calf serum (Hy-

Clone). Me180 and SiHa cervical carcinoma cell lines were grown in RPMIwith 0.5% or 10% Cosmic calf serum. BUD8 calls were grown in DMEMwith 10% FCS(HyClone). A43l-P cells weretransfected with StatiY7Ol-For a control vector, RCCMV (InV’ttrogen), using standard procedures withcalcium phosphate (45). Stable clones were selected for in DMEM wfth10% Cosmic calf serum containing 200 pg/mI of 6418 sulfate (Geneticin;Ufe Technologies, Inc.). Growth rate analysis and thymidine uptake stud-

ies were performed essentially as descnbed previously (40). Cells wereplated in six-well dishes at 20-30% confluence(typically 5 x 10�-5 x io�cells/well, depending on the cell type), permitted to adhere, and then

treated with either no additional ligand or EGF (R&D Systems) at 20 n� orIFNy(a generous gift from Amgen) at 15 ng/mI. At various times, cells weretrypsinized and counted using a Coulter counter. Alternatively, cells werepulsed with [�HJthymidine (1 �CVml; 70 CVmmol) for 30 mm to 1 h. Thecells were washed and collected, and TCA-precipitable counts weredetermined for each sample.

Northern, Western, Immunoprecipltatlons, and EMMa. Total RNA

from untreated or EGF- or IFNy-treated cells was isolated using TRIzolreagent (Ufe Technologies, Inc.). RNAs were resolved on a denaturingformaldehyde agarose gel (45), transferred to a nylon membrane (Zetap-robe; Bio-Rad), and probed with radiolabeled (Random Primer; Strat-

agene) IRFI and glyceraldehyde-3-phosphate dehydrogenase. Whole-cell extracts were isolated as descnbed previously (46). Nuclear extractswere isolated as described previously (47, 48). EMSAS were performed

essentially as described previously (49, 50) using a radiolabeled, m67double-stranded probe (dGATTTCCCGTAAATCAT; the consensus statbinding site is underlined) and nuclear extracts. lmmunoprecipitations

were carried out by incubating, at 4#{176}C,100-200 pg of total protein from

whole-cell extracts with appropriate antisera: EGFR-528, CDK2(M2),ERK1 , ERK2 (Santa Cruz Biotechnology), and agarose A beads (Once-gene Science) for several hours, followed by several washes in whole-cell

extract buffer. Phosphotyrosine blots were carried out using PY2O (Trans-duction Laboratories). lmmunoblots were visualized using ECL (DuPont

NEN, LJfe Sciences). CDK2 assays were performed essentially as de-scnbed (8, 40). MAPK assays were performed as for the CDK2 assaysubstituting MBP (Sigma)for histone Hi . Radiolabeled MBP was resolvedon a 12% polyacrylamide gel, transferred to nitrocellulose, and visualizedby autoradiography. For both CDK2 and MAPK assays, one-half of the

immunoprecip’rtated products were resolved on 8% gels transferred tonitrocellulose and probed with either CDK2 or ERK1 and ERK2 to assess

variability in immunoprecipitation and loading. Stati antisera (51) to the

COOH terminus of the proteln was used for supershifting Stati -containingDNA complexes. Stat3 antisera (17) was used as control serum. TheFlag-tagged StatlY-701-F (46) was produced by PCR and is subclonedinto the Apa/NotI sites of RCCMV (InVitrogen).

AcknowledgmentsWe thank Lois Cousseau for preparing the manuscript.

References1 . UlInch, A., and Schlessinger, J. Signal transduction by receptors withtyrosine kinase activity. Cell, 61: 203-212, 1990.

2. Gill, N. G., and Lazar, C. S. Increased phosphotyrosine content andinhibition of proliferation in EGF-treated A431 cells. Nature (Lond.), 293:305-307, 1981.

3. Barnes, D. W. Epidermal growth factor inhibits growth of A43l humanepidermoid carcinoma in serum-free cell culture. J. Cell Biol., 93: 1-4,

1982.

4. Kawamoto, T., Mendelsohn, J., Le, A., Sate, G., Lazar, C., and Gill, G.Relation of epidermal growth factor receptor concentration to growth ofhuman epidermoid carcinomaA43l cells. J. Biol. Chem., 259: 7761-7766,

1984.

5. Filmus, J., Pollack, M. N., Cailleau, R., and Buick, R. N. MDA468, ahuman breast cancer cell line with a high number of EGF receptors, hasan amplified EGF receptor gene and is growth inhibited by EGF. Biochem.Biophys. Res. Commun., 128: 898-905, 1985.

6. Davidson, N. E., Gelman, E. P., LJppman, M. E., and Dickson, R. B.Epidermal growth factor receptor gene expression in estrogen receptor

positive and negative human breast cancer cell lines. Mol. Endocrinol., 1:

216-223, 1987.

7. Gross, M. E, Zorbas, M. A., Daniels, V. J., Garcia, R., Gallick, G. E.,

Olive, M., Brattain, M. G., Boman, B. M., and Yeoman, L C. Cellulargrowth response to epidermal growth factor in colon carcinoma cells with

an amplified epidermal growth factor receptor derived from a familialadenomatous polyposis patient. Cancer Res., 51: 1452-1459, 1991.

8. Fan, Z., Lu, V., Wu, X., DeBlasio, A., Koff, A., and Mendelsohn, J.

Prolonged induction of p2lCipl/WAF1/CDK2/PCNA complex by epider-mal growth factor receptor activation mediates ligand-induced A431 cellgrowth inhibition. J. Cell Biol., 131: 235-242, 1995.

9. Chin, V. E, Kitagawa, M., Su, W. C., You, 1 H., Iwamoto, Y., and Fu,x. Y. Cell growth arrest and induction of cyclin-dependent kinase inhibitorp21 WAF1/CIP1 mediated by STAT1 . Science (Washington DC), 272:

719-722, 1996.

10. El-Deiry, W. S., Tokino, T., Velculescu, V. E., Canman, C. E., Levy,

D. B., Parsons, R., Trent, J. M., tin, W. E., Mercer, W. E., Kinzler, K. W.,

and Vogelstein, B. WAF1 , a potential mediator of p53 tumor suppression.Cell, 75: 817-825, 1993.

1 1 . El-Deiry, W., Harper, J. W., O’Connor, P. M., Velculescu, V. E.,

Canman, C. E., Jackman, J., Pietenpol, J., Burrell, M., Hill, D., Wang, Y.,Wiman, K., Mercer, W. E., Kastan, M., Kohn, K., Elledge, S., Kinzler, K.,and Vogelstein, B. WAF1/CIP1 is induced in p53-mediated G1 arrest andapoptosis. Cancer Res., 12: 1 169-1 174, 1994.

1 2. Tanaka, N., Ishihara, M., Lamphier, M. S., Nozawa, H., Matsuyama,T., Mak, T., Aizawa, S., Tokino, T., Oren, M., and Taniguchi, T. Cooper-ation of the tumour suppressors IRF1 and p53 in response to DNAdamage. Nature (Lond.), 382: 816-818, 1996.

13. Schlessinger, J. Cellular signaling by receptor tyrosine kinases. Ha’-

vey Lect., 89: 105-123, 1993-1994.

14. Lemmon, M. A., and Schlessinger, J. Regulation of signal transduc-

tion and signal diversity by receptor oligomerization. Trends Biochem.Sci., 19: 459-463, 1994.

15. Schlessinger, J. SH2JSH3 signaling proteins. Curr. Opmn. Genet. Dev.,

4: 25-30, 1994.

16. Fu, X-Y., and Zhang, J-J. Transcription factor p91 interacts with theepidermal growth factor receptor and mediates activation of the c-foegene promoter. Cell, 74: 1135-1 145, 1993.

17. Zhong, Z., Wen, Z., and Damell, J. E., Jr. Stat3: a STATfamily memberactivated by tyrosine phosphorylation in response to epidermal growthfactor and interleukin-6. Science (Washington DC), 264: 95-98, 1994.

18. Shual, K., Ziemiecki, A., WiIks, A. F., Harpur, A. G., Sadowski, H. B.,

Gilman, M. Z., and Damell, J. E., Jr. Polypeptide signaling to the nucleusthrough tyrosine phosphorylation of JAK and STAT proteins. Nature(Lond.), 366: 580-583, 1993.

19. Silvennoinen, 0., Schindler, C., Schlessinger, J., and Levy, D. E.Ras-independent growth factor signaling by transcription factor tyrosinephosphorylation. Science (Washington DC), 261: 1736-1739, 1993.

20. Lamer, A. C., David, M., Feldman, G. M., Igarashi, K-I., Hackett, R. H.,Webb, 0. S. A., Sweitzer, S. M., Petricon, E. F., III, and Finbloom, D. S.Tyrosine phosphorylation of DNA binding proteins by multiple cytokines.

Science (Washington DC), 261: 1730-1733, 1993.

21 . Damell, J. E., Jr., Kerr, I. M., and Stark, G. M. Jak-STAT pathways andtranscriptional activation in response to IFNs and other extracellular sig-naling proteins. Science (Washington DC), 264: 1415-1421 , 1994.

22. DameIl, J. E., Jr. STAT molecules and gene regulation. Science(Washington DC), 277: 1630-1635, 1997.

23. Leaman, 0. W., Pisharody, S., Flickinger, T. W., Commane, M. A.,

Schlessinger, J., Kerr, I. M., Levy, D. E., and Stark, G. R. Roles of JAKs inactivation of STATs and stimulation ofc-fos gene expression by epidermalgrowth factor. Mol. Cell. Biol., 16: 369-375, 1996.

24. David, M., Wong, L, Flavell, R., Thompson, S. A., Wells, A., Lamer,A. C., and Johnson, G. R. Stat Activation by EGF and amphiregulin-requirement for the EGF receptor kinase but not for tyrosine phosphoryl-

ation sites or Jakl . J. Biol. Chem., 271: 9185-9188, 1996.

25. Brown, C. L, Rubin, M., Masui, H., and Mendelsohn, J. Growth

inhibition by anti-EGF receptor monoclonal antibody in squamous cervical

512 Statl and EGF-induced Growth Arrest

carcinoma cells expressing TGF-a. Proc. Am. Assoc. Cancer Res., 35:381, 1994.

26. Hollenberg, M. D., and Cuatrecasas, P. Insulin and epidermal growthfactor Human fibroblast receptors related to deoxyribonucleic acid syn-thesis and amino acid uptake. J. Biol. Chem., 250: 3845-3853, 1975.

27. Carpenter, G., Lembach, K. J., Morrison, M. M., and Cohen, S.Characterization of the binding of 125-I-labeled epidermal growth factorto human fibroblasts. J. Biol. Chem., 250: 4297-4304, 1975.

28. Sadowski, H. B., Shuai, K., Damell, J. E., Jr., and Gilman, M. 1 Acommon nuclear signal transduction pathway activated by growth factorand cytokine receptors. Science (Washington DC), 261:1739-1744,1993.

29. Shual, K., Schindler, C., Prezioso, V. R., and Damell, J. E., Jr. Acti-vation of transcription by IFN-’y tyrosine phosphorylation of a 91 kD DNA

binding protein. Science (Washington DC), 259: 1808-1812, 1992.

30. Shuai, K., Horvath, C. M., Tsai-Huang, L H., Qureshi, S., Cowbum,

D., and Damell, J. E., Jr. Interferon activation of the transcription factorStat9i involves dimerization through SH2-phosphotyrosyl peptide inter-

actions. Cell, 76: 821-828, 1994.

31 . Sadowski, H. B., and Gilman, M. Z. Cell-free activation of a DNA-

binding protein by epidermal growth factor. Nature (Lond.), 362: 79-83,

1993.

32. Pine, R., Canova, A., and Schindler, C. Tyrosine phosphorylated p91binds to a single element in the ISGF2IIRF1 promoterto mediate inductionby IFN a and IFN ‘y, and is likely to autoregulate the p91 gene. EMBO J.,13: 158-167, 1994.

33. Muller, M., Laxton, C., Briscoe, J., Schindler, C., Improta, T., Damell,

J. E., Jr., Stark, G. A., and Kerr, I. M. Complementation of a mutant cellline: central role of the 91 -kDa polypeptide of ISGF3 in the interferon-a

and -‘y signal transduction pathway. EMBO J., 12: 4221-4228, 1993.

34. Carpenter, G., and Cohen, S. 125-I labeled human epidermal growthfactor. Binding, internalization and degradation in human fibroblasts.J. Cell Biol., 71: 159-171, 1976.

35. Carpenter, G., and Cohen, S. Human epidermal growth factor and theproliferation of human fibroblasts. J. Cell Physiol., 88: 227-237, 1976.

36. Masui, H., Castro, L, and Mendelsohn, J. Consumption of EGF by

A431 cells: evidence for receptor recycling. J. Cell Biol., 120: 85-93, 1993.

37. Qureshi, S. A, Leung, S., Kerr, I. M., Stark, G. R., and Damell, J. E.,Jr. Function of Stat2 protein in transcriptional activation by IFN-a. Mel.

Cell. Biol., 16: 288-293, 1996.

38. Nakajima, K., Yamanaka, Y., Nakae, K., Kojima, H., Ichiba, M., Kiuchi,N., Kitaoka, T., Fukada, T., Hibi, M., and Hirano, T. A central role for Stat3in IL-6 induced regulation of growth and differentiation in Ml leukemiacells. EMBO J., 15: 3651-3658, 1996.

39. Minami, M., lnoue, M., Wei, S., Takeda, M., Matsumoto, M., Kishi-moto, T., and Akira, S. Stat3 activation is a critical step in gpl3O mediated

terminal differentiation and growth arrest of a myeloid cell line. Proc. NatI.Acad. Sci. USA, 93: 3963-3966, 1996.

40. Bromberg, J., Horvath, C. M., Wen, 1, Schreiber, R. D., and Damell,J. E., Jr. Transcriptionally active Statl is required for the antiproliferatlve

effects of both interferon a and interferon gamma. Proc. Natl. Acad. Sd.

USA, 93: 7673-7678, 1996.

41 . Shuai, K., Liao, J., and Song, M. M. Enhancement of antiproliferative

activity of gamma interferon by the specific inhibition of tyrosine dephos-phorylation of Statl . Mol. Cell. Biol., 16: 4932-4941, 1996.

42. Su, W. C. S., Kitagawa, M., Xue, N. A., Xie, B., Garofalo, S., Cho, J.,Deng, C. X., Horton, W. A., and Fu, X. V. Activation of Statl by mutantfibroblast growth-factor receptor in thanatophoric dysplasia type II dwarf-sm. Nature (Lend.), 386: 288-292, 1997.

43. Coffer, P. J., and Kruijer, W. EGF receptor deletions define a regionspecifically mediating Stat transcription factor activation. Biochem. Bio-phys. Res. Commun., 210: 74-81, 1995.

44. Fukada, T., Hibi, M., Yamanaka, Y., Takahashi-Tezuka, M., Fujitani,Y., Tamaguchi, T., Nakajima, K., and Hirano, T. Two signals are necessary

for cell proliferation induced by a cytokine receptor gpl3O: involvement ofStat3 in anti-apoptosis. Immunity, 5: 449-460, 1996.

45. Ausubel, F., Brent, R., Kingston, R. E, Moore, D., Seidman, J. G.,Smith, J. A., and Struhl, K. Current Protocols in Molecular Biology, Vol. 1,

Sect. 4 and 9. New York: John Wiley and Sons, 1994

46. Shuai, K., Stark, G. R., Kerr, I. M., and Damell, J. E., Jr. A singlephosphotyrosine residue of Stat9l required for gene activation by inter-feron-’y. Science (Washington DC), 261: 1744-1746, 1993.

47. wen, 1, Thong, 1, and Damell, J. E J. Maximal activation of tran-scription by Statl and Stat3 requires both tyrosine and serine phospho-rylation. Cell, 82: 241-250, 1995.

48. Haspel, R. L, Salditt-Georgleff, M., and Damell, J. E., Jr. The rapidinactivation of nuclear tyrosine phosphorylated Stati depends upon a

protein tyrosine phosphatase. EMBO J., 15: 6262-6268, 1996.

49. Levy, D. E., Kessler, D. S., Pine, R. I., and Damell, J. E., Jr. Cytoplas-mic activation of ISGF3, the positive regulator of interferon-a stimulatedtranscription, reconstituted in vitro. Genes Dev., 3: 1362-1372, 1989.

50. Fried, M., and Crothers, D. M. Equilibria and kinetIcs of Inc repressor-operator interactions by polyacrylamide gel electrophoresls. Nucleic Ac-ide Res., 9: 6505-6525, 1981.

51. Schindler, C., Shuai, K., Pmezioso, V. R., and Damell, J. E., Jr.Interferon-dependent tyrosine phosphorylatlon of a latent cytoplasmictranscription factor. Science (Washington DC), 257: 809-815, 1992.